Participatory Procedural Justice in Performance Management: Reconstructing Goal Setting and Continuous Feedback

Conventional performance management systems often fail to secure procedural justice because goal setting and continuous feedback are treated as terminal managerial judgments rather than contestable procedural artifacts. We propose a two-tier architecture that reconstitutes these mechanisms as a participatory bias deliberation system. The operational tier retains the existing goal creation, progress check-in, rating generation, and calibration functions as a data-generating substrate. The deliberative tier, instantiated as a rotating Performance Council, audits the procedural justice properties of that substrate. Specifically, a transformer-based clarity scorer evaluates goal ambiguity, while a composite feedback-quality vector measures specificity, actionability, and temporal density. Demographic disparity analysis over these scores triggers a protocolized negotiation dialogue when a council-defined threshold is exceeded. This dialogue proceeds through three phases: surfacing silenced voice concerns, reason-giving, and root-cause mapping via a Bayesian network. The council then co-designs procedural safeguards—goal-clarity contracts, feedback scripts, rotating calibration panels, bias-incident reviews, and appeal routes—that are injected back into the operational tier. The integration is iterative and bidirectional, hence the conventional module is not merely augmented but procedurally reconstructed. The main contribution is a closed-loop mechanism that embeds procedural justice into the operational fabric of talent management. The proposed system offers a theoretically grounded and computationally specified alternative to static diversity training or post hoc bias audits, and it has significant implications for the design of auditable, participatory, and demographically equitable performance management.